Old Photo Restoration Brief
Plan a restoration of a damaged photograph. Preserve the person. Do not beautify them into someone else. SOURCE Era guess: [DECADE] Damage: [CREASES | STAINS | FADED | TORN CORNER | MISSING PIECE | MOLD] People: [COUNT AND WHO MATTERS] What I will accept being reconstructed: [BACKGROUND ONLY | SMALL TEAR | A MISSING HAND — BE EXPLICIT] What must not be invented: [FACES, CLOTHING PATTERNS, JEWELRY, ANY TEXT] Colorize: [NO | YES, CONSERVATIVE] Output: [PRINT AT SIZE | SCREEN] Reference photos of the same people, if any: [YES | NO] PLAN 1) Digitize notes: if I only have a phone snap of a print, say whether to re-scan and at what resolution before any generation. 2) Order of operations: dust and scan cleanup, geometry (crease bulge), tonal rebuild, then any generative fill. Explain why fill comes last. 3) For each damaged region: restore by cloning from nearby original texture versus generative fill. Faces and eyes default to manual reconstruction from remaining detail. If the face is gone and I have no reference, stop and say the honest version is a tasteful crop or an unrepaired area — do not synthesize a new identity. 4) Grain: match the original film grain instead of delivering plastic skin. 5) Colorization, if requested: a restrained palette for the era, with a note that colors are an interpretation, not a recovery. 6) File handling: work on a copy, keep the original, name the layers, export a before/after at the same crop for the family. 7) A short caption I can put in an album: what was repaired and what was left, in two sentences. No beauty retouch, no slimming, no opened eyes that were closed, no removed people.
🌟 Example Output / Preview
Prompt Metadata
Primary Use Cases:
- •E-commerce product visual polishing
- •Brand media asset workflow enhancement
- •High-end retouching & cinematic grade balancing
Associated Tags:
💡 Pro Tips & Advice
1. Use bracketed items: Be sure to fill out all [PLACEHOLDER] elements with specific details before sending the prompt to the AI model.
2. Adjust temperature: For creative tasks, set AI temperature higher (e.g., 0.8), or lower (e.g., 0.2) for strict coding/technical tasks.
🔗 Related AI Prompts
Background removal instructions
I need to remove the background from this image of a [subject]. Provide precise instructions for a designer on how to mask around ...
Upscaling low-res images
Act as a photo editor. Recommend the best workflow and AI tools (like Topaz or Magnific) to upscale a low-resolution, heavily comp...
Color grading to brand palette
I have a set of photos that need to be color graded to match our brand's aesthetic. Our brand colors are [color hex codes], and ou...